基于多尺度子带划分的雾霾天气图像增强方法

Foggy and lowery image enhancement method based on multi-scale and sub-band division

目前一些城市马路和高速公路因为雾霾天气原因导致获取的图像清晰度和对比度较低,对行车安全性和道路监控的可靠性都产生了不良的影响。针对这类图像的特点,提出一种基于多尺度子带划分的阈值处理方法。与其他图像增强方法不同的是,这种方法在小波变换多尺度分析的理论基础上,设计出了在不同尺度和不同方向子带图像上同时具有自适应性的小波系数阈值处理方法,通过这种方法重构出的目标图像能够很好地改善图像整体视觉效果。

Because of foggy and lowery atmosphere, the images acquired from some roads in cities and highways have a relatively low resolution and contrast. So whether from the road traffic safety or the reliability of the road monitoring, it has had been an adverse impact. On the characteristics of this type of images, a threshold processing method based on multi-scale and sub-band division was proposed, which is different from the traditional method for image enhancement. It designed a new seff- adaptive threshold processing method that can process the wavelet coefficients of the different scales and sub-band images in the different directions on the basis of multi-scale analysis of wavelet transform. The object images reconstructed by this method can be well improved in the overall visual effect.